API Reference

The mcrnmf package exposes two primary modules:

1. models - contains core decomposition algorithms

2. nnls - provides a standalone solver for estimating concentration profiles \(H\) given fixed spectra \(W\)

models

The models module provides implementations of various Nonnegative Matrix Factorization (NMF) algorithms used in Multivariate Curve Resolution (MCR).

mcrnmf.models.FroALS	Frobenius norm-based Nonnegative Matrix Factorization (NMF) using Alternating Least Squares (ALS) method.
mcrnmf.models.FroFPGM	Frobenius norm-based Nonnegative Matrix Factorization (NMF) using Fast Projected Gradient Method (FPGM).
mcrnmf.models.MinVol	Minimum-Volume Nonnegative Matrix Factorization (NMF) implementation using Fast Projected Gradient Method.
mcrnmf.models.SNPA	Successive Nonnegative Projection Algorithm (SNPA) for NMF initialization.

nnls

The nnls module offers a fast projected gradient solver to estimate \(H\) from new data \(X\) when \(W\) is fixed and known.

mcrnmf.nnls.FPGM

An implementation of Non-negative Least Squares solver using Fast Projected Gradient Method.